Back

Maybe this is beating a dead horse, but I don't think you can really talk about Matrix without discussing the original sin of its branding.

Nearly every component of the system has been renamed or relaunched multiple times over the project's nine years. Element vs Element X feels like a minor problem after the era of Vector vs Riot vs Element. Synapse has actually held on to its name, but had a complete replacement (Dendrite) launched and then badly stagnated.

I don't know exactly how to explain this, it's very weird organizational behavior. Too much leadership by engineering? Too little leadership in general? but at least among people I know that have been long-term Matrix users, it has done a LOT to deteriorate trust in the system. New Vector has nearly a Google-level reputation for abandoning projects.

To be fair less has been abandoning than rebranding, but as a user those rebrands have often been painful and confusing processes.

Things to think about when building a system: user-facing consistency is important for trust and ease of use. administrator-facing consistency is important for stability and maintainability.

You shouldn't rebrand something unless you truly have no other choice, and I'm not convinced that even the political issues around the "Riot" name met that bar.

You shouldn't launch a complete rewrite of infrastructure unless you really intend to follow through on it and end development of the old one.

And I know everyone will say "but it's an open system, the fact that there are multiple clients is a feature."

But, first, non-Element clients have such a small userbase that they barely matter from the perspective of the project.

Second, the constant churn in product direction for both the protocol and the de facto reference client have massively contributed to the reliability and interoperability problems of third-party clients.

I might put it this way: having multiple clients and an open system is absolutely a feature. But if the revenue-backed core project can barely maintain a single client, the chances of the community developing a more stable and complete alternative are close to nil.

Open-source projects cannot rely on their community to fill a void in product leadership. The best case, in that scenario, is a fork, and that's a very very bad case for real users.

I sincerely like Matrix and I want it to succeed, I appreciate the work of New Vector and have generally had positive interactions with employees there. I don't want to be too mean to anyone.

But if Matrix is going to "take off" on the scale of Slack or FB Messenger or whatever goal you might set, it badly needs more consistent direction and more investment in UX. As a consumer technology, it is just not there. It's getting closer! but it's not there, and progress is alarmingly slow.

From my outsider perspective I see two root problems:

1) New Vector gets its revenue from large institutional users and feels like it is too busy chasing that revenue to care that much about noninstitutional users. This is of course a classic problem with services-funded FOSS.

2) Matrix as an open project spent a long time very interested in adding new features, but adding features greatly outpaced implementing them in Element, so Element itself has routinely had compatibility problem.